Similar Tasks Research Articles

Oxyanion Engineering on RuO2 for Efficient Proton Exchange Membrane Water Electrolysis

In proton exchange membrane water electrolysis (PEMWE), the anode oxygen evolution reaction (OER) catalysts rely heavily on the expensive and scarce iridium‐based materials. Ruthenium dioxide (RuO2) with lower price and higher OER activity, has been explored for the similar task, but has been restricted by the poor stability. Herein, we developed an anion modification strategy to improve the OER performance of RuO2 in acidic media. The designed multicomponent catalyst based on sulfate anchored on RuO2/MoO3 displays a low overpotential of 190 mV at 10 mA cm‐2 and stably operates for 500 hours with a very low degradation rate of 20 μV h‐1. When assembled in a PEMWE cell, this catalyst as an anode shows an excellent stability at 500 mA cm‐2 for 150 h. Experimental and theoretical results revealed that MoO3 could stabilize sulfate anion on RuO2 surface to suppress its leaching during OER. Such MoO3‐anchored sulfate not only reduces the formation energy of *OOH intermediate on RuO2, but also impedes both the surface Ru and lattice oxygen loss, thereby achieving the high OER activity and exceptional durability.

Examining the relationship between psychosocial adversity and inhibitory control: an fMRI study of children growing up in extreme poverty.

Exposure to psychosocial adversity (PA) is associated with poor behavioral, physical, and mental health outcomes in adulthood. As these outcomes are related to alterations in developmental processes, growing evidence suggests that deficits in executive functions-inhibitory control in particular-may, in part, explain this relationship. However, literature examining the development of inhibitory control has been based on children in higher resource environments, and little is known how low resource settings might exacerbate the link between inhibitory control and health outcomes. In this context, we collected fMRI data during a Go/No-Go inhibitory control task and PA variables for 68 children 5 to 7 years of age living in Dhaka, Bangladesh, an area with a high prevalence of PA. The children's mothers completed behavioral questionnaires to assess the child's PA and their own PA. Whole-brain activation underlying inhibitory control was examined using the No-Go versus Go contrast, and associations with PA variables were assessed using whole-brain regressions. Childhood neglect was associated with weaker activation in the right posterior cingulate, whereas greater family conflict, economic stress, and maternal PA factors were associated with greater activation in the left medial frontal gyrus, right superior and middle frontal gyri, and left cingulate gyrus. These data suggest that neural networks supporting inhibitory control processes may vary as a function of exposure to different types of PA, particularly between those related to threat and deprivation. Furthermore, increased activation in children with greater PA may serve as a compensatory mechanism, allowing them to maintain similar behavioral task performance.

Improved keypoint localization network for tea bud based on YOLO framework

The precise localization of picking points is the primary challenge that should be addressed by the robotic tea bud harvester. The efficient identification and accurate positioning of picking points directly affects the efficiency of tea buds harvesting. Compared with the target detection and semantic segmentation algorithms, the keypoint detection algorithms do not require to incorporate target shape features for accurately localizing the picking point, which allows to significantly decrease the computational load, and thus facilitates the subsequent point cloud processing. This study introduces a tea bud keypoint detection network referred to as TBKNet. This model uses the concept of structural reparameterization to reconstruct the CSPDarknet53 backbone network, and constructs an asymptotic parallel feature fusion network by incorporating the CA attention mechanism into the asymptotic feature pyramid network. In addition, in order to accelerate the convergence of the model, a keypoint regression loss function, which considers the angle and Euclidean distance, is proposed. The experimental results show that TBNKet reaches the highest accuracy with an average precision of 87.1 % for keypoint detection. It has 5.7 %, 16.9 %, and 3.3 % higher mAP values compared with the DEKR, YOLOPose, and YOLOPv8-Pose models, respectively. This is efficient in identifying and localizing keypoints of tea buds, and it can provide guidance for similar intensive harvesting tasks.

Effects of the Regular Use of Virtual Environments on Spatial Navigation and Memory.

Introduction: The cognitive effects of video games have garnered increasing attention due to their potential applications in cognitive rehabilitation and evaluation. However, the underlying mechanisms driving these cognitive modifications remain poorly understood. Objectives: This study investigates the fundamental mnemonic processes of spatial navigation, pattern separation, and recognition memory, closely associated with the hippocampus. Our objective is to elucidate the interaction of these cognitive processes and shed light on rehabilitation mechanisms that could inform the design of video games aimed at stimulating the hippocampus. Method: In this study, we assessed 48 young adults, including both video game players and non-players. We utilized virtual reality and cognitive tasks such as the Lobato Virtual Water Maze and the Mnemonic Similarity Task to evaluate their cognitive abilities. Results: Our key findings highlight that gamers exhibit heightened pattern separation abilities and demonstrate quicker and more accurate spatial learning, attributed to the cognitive stimulation induced by video games. Additionally, we uncovered a significant relationship between spatial memory, guided by environmental cues, and pattern separation, which serves as the foundation for more efficient spatial navigation. Conclusions: These results provide valuable insights into the cognitive impact of video games and offer potential for monitoring changes in rehabilitation processes and early signs of cognitive decline through virtual reality-based assessments. Ultimately, we propose that examining the relationships between cognitive processes represents an effective method for evaluating neurodegenerative conditions, offering new possibilities for early diagnosis and intervention.

EXAMINING THE ROLE OF CONTEXT IN STATISTICAL LITERACY OUTCOMES USING AN ISOMORPHIC ASSESSMENT INSTRUMENT

The Guidelines for Assessment and Instruction in Statistics Education (GAISE) College Report advocates for use of real data with context and purpose. This work contributes to the growing literature on assessing statistical literacy by investigating the influence of context as it relates to assessment performance among post-secondary introductory statistics students. We discuss the development of an isomorphic form of an existing assessment instrument, and report results which concluded that test takers demonstrated lower statistical literacy scores when assessment tasks incorporated real data from published studies as context when compared with functionally similar tasks such as those with a contrived data set and a realistic context.

Ant and termite collective behavior: Group‐level similarity arising from individual‐level diversity

AbstractAnimal collective behavior produces group‐level patterns that emerge from individual‐level interactions among members. Such sophisticated behavioral coordination can be observed across various taxa, indicating evolutionary convergence. Among the most striking examples are two social insect lineages: ants and termites. Ants and termites evolved socially independently, inhabit terrestrial and subterranean environments, and face a similar collective task of maintaining their colonies with many individuals. This results in the convergent evolution of collective phenomena, such as constructing complex nest structures or movement coordination based on chemical and tactile cues. However, although the striking similarity attracts the most attention, little research has focused on how ants and termites differ in collective behaviors at individual and group levels. Here, we review the similarities and distinctions of collective behaviors in ants and termites. Even with similarities in group‐level patterns, rich diversity exists in mechanisms of behavioral coordination and the functions of collective patterns. Comparative analysis of collective behavior is challenging, but recent advances in automatic movement tracking methodologies have greatly expanded the potential for generating further insights. Finally, we conclude that comparative collective behavioral analysis can enhance our ability to understand biodiversity and provide alternative solutions for collective problem‐solving in many fields.

Typology of professional trajectories of gifted individuals using neural network analysis

Based on a data set (100 biographies) created by the authors through content analysis of biographical material about outstanding scientists of the 19th and 20th centuries in the humanities and natural sciences, the clustering of professional trajectories of gifted individuals was carried out. Neural network analysis based on self-organizing Kohonen maps was used as a clustering method. The professional trajectories were formed within the framework of the behavioral model of the linear-stage approach to studying life cycles. Within this approach, career and professional self-realization are understood as a sequence of evolutionary stages fixed in their order of occurrence. Each stage was encoded, and the biographies were transformed into a vector system. In turn, the task of clustering consisted in dividing a hundred vectors into typical groups with several real-valued coordinates. The criteria for the quality of clustering were the minimum sum of quantization errors and the silhouette coefficient. As a result of the study, seven professional trajectories of gifted individuals were identified and interpreted. The analysis of trajectories was carried out from the point of view of the speed of success (average age of success) and those factors and conditions of the life path that could affect either rapid or slow achievement of professional goals and self-realization. This example demonstrates the possibilities and limitations of using neural network analysis for solving similar research tasks, especially when working with complex cluster forms and finding their optimal number.

Simple Data Transformations for Mitigating the Syntactic Similarity to Improve Sentence Embeddings at Supervised Contrastive Learning

Contrastive learning of sentence representations has achieved great improvements in several natural language processing tasks. However, the supervised contrastive learning model trained on the natural language inference (NLI) dataset is insufficient to elucidate the semantics of sentences since it is prone to make a prediction based on heuristics. Herein, by using the ParsEVAL and the word overlap metric, it is shown that sentence pairs in the NLI dataset have strong syntactic similarity and propose a framework to compensate for this problem in two aspects. 1) Apply simple syntactic transformations to the hypothesis and 2) expand the objective to SupCon Loss to leverage variants of sentences. The method is evaluated on semantic textual similarity (STS) tasks and transfer tasks. The proposed methods improve the performance of the BERT‐based baseline in STS Benchmark and SICK Relatedness by 1.48% and 2.2%. Furthermore, the model achieves 82.65% on the HANS benchmark dataset, to the best of our knowledge, which is a state‐of‐the‐art performance demonstrating that our approach is effective in grasping semantics without heuristics in the NLI dataset at supervised contrastive learning. The code is available at https://github.com/whnhch/Break-the-Similarity.

Two aspects of feed-forward control of action stability: effects of action speed and unexpected events.

We explored two types of anticipatory synergy adjustments (ASA) during accurate four-finger total force production task. The first type is a change in the index of force-stabilizing synergy during a steady state when a person is expecting a signal to produce a quick force change, which is seen even when the signal does not come (steady-state ASA). The other type is the drop in in the synergy index prior to a planned force change starting at a known time (transient ASA). The subjects performed a task of steady force production at 10% of maximal voluntary contraction (MVC) followed by a ramp to 20% MVC over 1s, 3s, and as a step function (0s). In another task, in 50% of the trials during the steady-state phase, an unexpected signal could come requiring a quick force pulse to 20% MVC (0-surprise). Inter-trial variance in the finger force space was used to quantify the index of force-stabilizing synergy within the uncontrolled manifold hypothesis. We observed significantly lower synergy index values during the steady state in the 0-ramp trials compared to the 1-ramp and 3-ramp trials. There was also larger transient ASA during the 0-ramp trials. In the 0-surprise condition, the synergy index was significantly higher compared to the 0-ramp condition whereas the transient ASA was significantly larger. The finding of transient ASA scaling is of importance for clinical studies, which commonly involve populations with slower actions, which can by itself be associated with smaller ASAs. The participants varied the sharing pattern of total force across the fingers more in the task with "surprises". This was coupled to more attention to precision of performance, i.e., inter-trial deviations from the target as reflected in smaller variance affecting total force, possibly reflecting higher concentration on the task, which the participants perceived as more challenging compared to a similar task without surprise targets.

Event-related potentials during encoding coincide with subsequent forced-choice mnemonic discrimination

Computational models and eye-tracking research suggest that encoding variability accounts for the reduced recognition of targets (A) when paired with non-corresponding lures (B′) relative to corresponding lures (A′). The current study examined whether neural activity during learning coincided with subsequent performance on the forced-choice Mnemonic Similarity Task (MST). Event-related potential responses were collected during encoding while young adults completed A–B′ and A–A′ trials of the forced-choice MST. Consistent with previous research, performance was lower on A–B′ trials than A–A′ trials. The subsequent memory effect was not significant for the A–A′ test format. However, for A–B′ trials, we observed a significant Accuracy × Stimulus interaction 1000–1200 ms poststimulus onset across frontal and fronto-central electrodes. As hypothesized, subsequently correct A–B′ trials were associated with a larger amplitude response at encoding to the target (A) than the original version of the non-corresponding lure (B). However, subsequently incorrect trials were associated with a larger amplitude response to the non-corresponding lure (B) than the target stimulus (A). These findings provide additional support for the effect of encoding variability on mnemonic discrimination.

Learning from NN-based extended PID control for a class of high-order uncertain nonlinear systems

As is well known, traditional PID controllers are the most widely used type of controllers in industrial control due to their simplicity and other advantages. However, the existing research results on PID controllers lack a general theory for addressing the nonlinearity and uncertainty present in practical systems. In the era of intelligent control, it is highly valuable to design an intelligent PID controller with learning capability for complex unknown nonlinear systems. In this paper, we consider an extended form of PID controller based on deterministic learning (DL) for high-order nonlinear systems with unknown dynamics. The introduction of neural networks (NNs) provides an effective tool to compensate for the composite unknown nonlinearities in controlled systems, overcoming the limitations of traditional PID controllers relying solely on feedback mechanisms to combat unknown nonlinearities. Based on the Lyapunov theory, it not only ensures the stability and tracking control of the system but also simplifies the selection of controller parameters. Furthermore, based on the DL theory, the composite unknown nonlinear dynamics can be accurately compensated and approximated, and the acquired knowledge can be stored in a constant-value neural network. When encountering similar control tasks, the acquired knowledge can be quickly accessed to construct a knowledge-based extended PID controller, thereby improving the overall control performance. Simulation results demonstrate the effectiveness of the proposed algorithm.

It’s about time: mitigating cancer-related cognitive impairments through findings from computational models of the Wisconsin Card Sorting Task

BackgroundMany cancer survivors experience cancer-related cognitive impairment (CRCI), often with significant negative consequences across various life domains. Emerging evidence suggests that allowing additional time to process information before acting may be a useful strategy for those with CRCI to mitigate some of its impacts. The Wisconsin Card Sorting Task (WCST), a measure of general cognition, has shown that for some cancer survivors, longer task completion time facilitates similar task performance outcomes to control populations concerning perseveration errors; a key performance metric of the WCST. However, assessing if this strategy may be useful, as well as determining for whom it may be useful, with regard to strengths and weaknesses among select cognitive domains, is challenging due to factors such as the problem of task impurity. Accordingly, this study provides an initial computational and experimental assessment of whether additional time to process information before acting is a useful strategy for those with CRCI.MethodsWe simulated individual cognitive differences observed in humans by varying contributions of executive functioning components (updating, shifting, inhibition) to yield 48 distinct computational models of the WCST. Our main manipulation was then to provide these models with more or less time (at three levels of 20, 40 and 60 cycles) before models executed an action to sort a given card. We compared the number of perseveration errors on the WCST produced by the computational models. Additionally, we determined models that simulated the performance of cancer survivors on the WCST by comparing the number of perseveration errors produced by the models to human data.ResultsAdditional processing time resulted in the models producing significantly fewer perseveration errors, supporting our hypothesis. In addition, 8 unique models simulated the performance of cancer survivors on the WCST. Additional time appeared to have a positive influence on performance primarily by mitigating the impacts of severe inhibition impairments. For more severe global executive function impairments, a substantial amount of additional time was required to mitigate the impacts of the impairments. For the most severe impairments, additional time was unable to adequately mitigate the impact on performance.ConclusionAdditional processing time may be a useful strategy to rectify perseveration errors among cancer survivors with CRCI. Our findings have implications for the development of practical strategies, such as workload and deadline management in occupational settings, which may mitigate the negative effects of CRCI.

Oscillatory Neural Correlates of Police Firearms Decision-Making in Virtual Reality.

We investigated the neural signatures of expert decision-making in the context of police training in a virtual reality-based shoot/don't shoot scenario. Police officers can use stopping force against a perpetrator, which may require using a firearm and each decision made by an officer to discharge their firearm or not has substantial implications. Therefore, it is important to understand the cognitive and underlying neurophysiological processes that lead to such a decision. We used virtual reality-based simulations to elicit ecologically valid behavior from authorized firearms officers (AFOs) in the UK and matched novices in a shoot/don't shoot task and recorded electroencephalography concurrently. We found that AFOs had consistently faster response times than novices, suggesting our task was sensitive to their expertise. To investigate differences in decision-making processes under varying levels of threat and expertise, we analyzed electrophysiological signals originating from the anterior cingulate cortex. In line with similar response inhibition tasks, we found greater increases in preresponse theta power when participants inhibited the response to shoot when under no threat as compared with shooting. Most importantly, we showed that when preparing against threat, theta power increase was greater for experts than novices, suggesting that differences in performance between experts and novices are due to their greater orientation toward threat. Additionally, shorter beta rebounds suggest that experts were "ready for action" sooner. More generally, we demonstrate that the investigation of expert decision-making should incorporate naturalistic stimuli and an appropriate control group to enhance validity.

Hybrid fault tolerant cost aware mechanism for scientific workflow in cloud computing

Cloud computing provides solutions for diverse commercial and academic applications which is the primary goal. Scientific workflows are used in the cloud-computing environment to analyses large-scale scientific applications. For scientific workflows, many data is required, and a single scientific workflow that includes hundreds of stages, depending on the application's time restrictions, task failures, money limits, incorrect task organization, and task management issues can all hinder the implementation of scientific methods. In light of this, a cloud-based scientific workflow management and scheduling system that is fault-tolerant and data-oriented method are proposed. This research designs a novel hybrid cost-aware fault tolerant (HCFT) mechanism for minimizing the cost. Moreover, HCFT integrates optimal clustering and efficient resource utilization through parallel and distributed execution. Novelty of HCFT lies in novel clustering of the similar task for improvisation, CyberShake, laser interferometer gravitational wave observatory (LIGO), Montage, and sRNA identification protocol using high throughput technology (SIPHT) processes are used in the simulations to evaluate the performance of the proposed approach.

Unveiling surgical expertise through machine learning in a novel VR/AR spinal simulator: A multilayered approach using transfer learning and connection weights analysis

BackgroundVirtual and augmented reality surgical simulators, integrated with machine learning, are becoming essential for training psychomotor skills, and analyzing surgical performance. Despite the promise of methods like the Connection Weights Algorithm, the small sample sizes (small number of participants (N)) typical of these trials challenge the generalizability and robustness of models. Approaches like data augmentation and transfer learning from models trained on similar surgical tasks address these limitations. ObjectiveTo demonstrate the efficacy of artificial neural network and transfer learning algorithms in evaluating virtual surgical performances, applied to a simulated oblique lateral lumbar interbody fusion technique in an augmented and virtual reality simulator. DesignThe study developed and integrated artificial neural network algorithms within a novel simulator platform, using data from the simulated tasks to generate 276 performance metrics across motion, safety, and efficiency. Innovatively, it applies transfer learning from a pre-trained ANN model developed for a similar spinal simulator, enhancing the training process, and addressing the challenge of small datasets. SettingMusculoskeletal Biomechanics Research Lab; Neurosurgical Simulation and Artificial Intelligence Learning Centre, McGill University, Montreal, Canada. ParticipantsTwenty-seven participants divided into 3 groups: 9 post-residents, 6 senior and 12 junior residents. ResultsTwo models, a stand-alone model trained from scratch and another leveraging transfer learning, were trained on nine selected surgical metrics achieving 75 % and 87.5 % testing accuracy respectively. ConclusionsThis study presents a novel blueprint for addressing limited datasets in surgical simulations through the strategic use of transfer learning and data augmentation. It also evaluates and reinforces the application of the Connection Weights Algorithm from our previous publication. Together, these methodologies not only enhance the precision of performance classification but also advance the validation of surgical training platforms.

ВЛИЯНИЕ ПОГРЕШНОСТЕЙ ИЗМЕРЕНИЯ КООРДИНАТ СУДНА НА КАЧЕСТВО ПРОЦЕССА УПРАВЛЕНИЯ

Autopilots are an integral part of modern ships. However, at present, many tasks are emerging, including the ones in the unmanned navigation field, requiring the use of new control algorithms and new means of monitoring navigation parameters. In particular, to solve such problems as keeping a vessel on a given trajectory, automating the ships’ divergence and similar tasks, it is necessary to have means of continuous monitoring of the ship’s location. For this purpose, receivers of global navigation satellite systems can be applied. To model control processes, it is necessary to know the distribution of positioning errors of modern devices. Some works show that over long time intervals (a day or more), positioning errors can be considered a random process with a Gaussian distribution. This paper presents the experimental study results of the distribution of positioning errors of a ship navigation receiver operating in the GPS and GLONASS systems, in a combined GPS/GLONASS mode, as well as in the indicated modes with differential corrections. The paper shows that at time intervals of seconds – minutes (characteristic of ship’s control systems), positioning errors of the navigation receiver cannot be represented by a random process with a Gaussian distribution law. The paper presents the results of a mathematical modelling of the process of holding a ship on a given trajectory using experimental values of positioning sensor errors. For the research, a mathematical model of an innovative ship with a wheel-propulsion steering complex is used. An algorithm is created that makes it possible to keep a vessel on a given trajectory with an accuracy acceptable for a practical use, commensurate with deviations determined by the design features of the ship’s hull and its propulsors.

Telepractice in Aphasia: Troubleshooting and Planning Assessments Using the Virtual-Appropriate Decision Approach

Purpose: This article describes the use of the Virtual-Appropriate Decision Approach (VADA). VADA is a systematic data gathering and problem-solving approach for planning and troubleshooting human, setting, and technology-related issues that may arise when conducting aphasia assessments via telepractice. Case examples are provided to illustrate and discuss its application for clinical and research settings. Method: We describe how VADA was used to analyze and mitigate challenges during administration of aphasia assessments in five (four cisgender males, one cisgender female) research participants via telepractice. Case discussions are organized by aphasia types and identify a variety of human, setting, and technology-related issues, which are analyzed via the VADA framework. Successes, challenges, and recommendations for conducting aphasia assessment via telepractice are included for each case. Conclusions: Unique challenges exist when using telepractice to conduct clinically valid standardized aphasia assessments for research or in the clinic. VADA is a systematic process to gather, analyze, evaluate, and address human, setting, and technology factors that impact aphasia assessments conducted via telepractice. The use of aphasia-friendly printed materials, assistance from communication partners, and similar task practice were adaptations that helped facilitate the assessment process.

Addressing environmental stochasticity in reconfigurable intelligent surface aided unmanned aerial vehicle networks: Multi-task deep reinforcement learning based optimization for physical layer security

Deep Reinforcement Learning (DRL) is a powerful tool for optimizing communications in reconfigurable intelligent surface (RIS)-assisted millimeter-wave unmanned aerial vehicle (UAV) systems, particularly for ensuring physical layer security by dynamically maximizing signal strength for legitimate users while minimizing leakage towards eavesdroppers. However, existing approaches often overfit the DRL agent to a non-stochastic environment, where the trajectories of users and eavesdroppers are fixed. This results in a single-task DRL agent that cannot generalize beyond the training environment/task. In response, this study introduces a novel multi-task DRL framework to address this limitation. Firstly, recognizing that not all tasks share the same solution, we proposed a reward-driven clustering method to group similar tasks based on their reward scores. This enables the training of a shared multi-task DRL agent for each cluster. Each cluster can be treated as an expertise, and the cluster-specific multi-task DRL agents are considered the 'experts'. Meanwhile, we leverage federated learning (FL) to speed up the training of the shared multi-task DRL agent. By combining these two techniques, our reward-driven clustered FL autonomously groups similar tasks together rapidly. Simulation results demonstrate a reduction in the number of agents from 50 to 10 specialized 'experts', while achieving a 10X speedup compared to conventional single-task DRL training. Furthermore, the study presents a mixture of experts (MoE) model that unifies these 'experts' into a single model. The MoE is trained to select the most suitable 'expert' when encountering new tasks. Our MoE exhibits robustness by effectively generalizing to 50 novel tasks without fine-tuning.

Majority voting of doctors improves appropriateness of AI reliance in pathology

As Artificial Intelligence (AI) making advancements in medical decision-making, there is a growing need to ensure doctors develop appropriate reliance on AI to avoid adverse outcomes. However, existing methods in enabling appropriate AI reliance might encounter challenges while being applied in the medical domain. With this regard, this work employs and provides the validation of an alternative approach – majority voting – to facilitate appropriate reliance on AI in medical decision-making. This is achieved by a multi-institutional user study involving 32 medical professionals with various backgrounds, focusing on the pathology task of visually detecting a pattern, mitoses, in tumor images. Here, the majority voting process was conducted by synthesizing decisions under AI assistance from a group of pathology doctors (pathologists). Two metrics were used to evaluate the appropriateness of AI reliance: Relative AI Reliance (RAIR) and Relative Self-Reliance (RSR). Results showed that even with groups of three pathologists, majority-voted decisions significantly increased both RAIR and RSR – by approximately 9% and 31%, respectively – compared to decisions made by one pathologist collaborating with AI. This increased appropriateness resulted in better precision and recall in the detection of mitoses. While our study is centered on pathology, we believe these insights can be extended to general high-stakes decision-making processes involving similar visual tasks.

Hierarchical representation learning for next basket recommendation

The task of next basket recommendation is pivotal for recommender systems. It involves predicting user actions, such as the next product purchase or movie selection, by exploring sequential purchase behavior and integrating users’ general preferences. These elements may converge and influence users’ subsequent choices. The challenge intensifies with the presence of varied user purchase sequences in the training set, as indiscriminate incorporation of these sequences can introduce superfluous noise. In response to these challenges, we propose an innovative approach: the Selective Hierarchical Representation Model (SHRM). This model effectively integrates transactional data and user profiles to discern both sequential purchase transactions and general user preferences. The SHRM’s adaptability, particularly in employing nonlinear aggregation operations on user representations, enables it to model complex interactions among various influencing factors. Notably, the SHRM employs a Recurrent Neural Network (RNN) to capture extended dependencies in recent purchasing activities. Moreover, it incorporates an innovative sequence similarity task, grounded in a k-plet sampling strategy. This strategy clusters similar sequences, significantly mitigating the learning process’s noise impact. Through empirical validation on three diverse real-world datasets, we demonstrate that our model consistently surpasses leading benchmarks across various evaluation metrics, establishing a new standard in next-basket recommendation.